RCM Studies to Enable Gasoline-Relevant Low Temperature ... · PDF fileGASOLINE-RELEVANT LOW...
Transcript of RCM Studies to Enable Gasoline-Relevant Low Temperature ... · PDF fileGASOLINE-RELEVANT LOW...
RCM STUDIES TO ENABLE GASOLINE-RELEVANT LOW TEMPERATURE COMBUSTION
drhgfdjhngngfmhgmghmghjmghfmf
S. SCOTT GOLDSBOROUGH (PI), ALEKSANDR FRIDLYAND, JEFFREY SANTNER, COLIN BANYON
08 JUNE 2016
Project ID # ACE054FY2016 DOE Vehicle Technologies Program Annual Merit ReviewAdvanced Combustion Engine R&D / Fuel and Lubricant TechnologiesWednesday, June 8, 2016Program Managers: Gurpreet Singh, Leo Breton
This presentation does not contain any proprietary, confidential or otherwise restricted information
OVERVIEW
Partners• ANL – Lead, Goldsborough (PI)• LLNL – gasoline surrogate model,
simulation tools• KAUST, Chevron – fuels, fuel models• UC Berkeley, SNL – HCCI engine data• Northeastern U. – UQ/GSA mechanism
diagnostics• International RCM Workshop
Barriers• Lack of fundamental knowledge
of advanced combustion engine regimes
• Lack of modeling capability for combustion and emission control
Timeline• Project started FY 2011• Project directions and continuation
reviewed annually, and in FY 2017 VTO Lab Call
Budget• Project funded by DOE / VTP→ FY 2014 funding: $325 k→ FY 2015 funding: $500 k→ FY 2016 funding: $490 k
2
OBJECTIVES AND RELEVANCE TO DOE
Acquire fundamental data, and help develop / validate / refinechemical kinetic and relevant models for transportation-relevant fuels(conventional and future gasolines, diesels and additives) at conditionsrepresentative of advanced combustion regimes, leveragingcollaborations with BES-funded groups, and researchers across thebroader community.
Predictive simulations with thesemodels, which require low associateduncertainties, could be utilized toovercome technical barriers to lowtemperature combustion (LTC), andachieve required gains in engineefficiency and pollutant reductions.
HO2+HO2=H2O2+O2doi:10.1021/jz400874s
3
PROJECT MILESTONESFY 2016
Task Milestone Status1 Acquire ignition delay measurements for gasoline surrogate
components: 5-member ring naphthenes (cyclopentane, methyl cyclopentane, ethyl cyclopentane)
2 Acquire additional ignition measurements for full boiling-range gasoline, and blends of this with ethanol: FACE-F (E0, E10, E20, E30, E100)
3 Acquire ignition measurements for multi-component surrogate blends to mimic ‘neat’, and ethanol-blended gasoline
FY16 – Q3
4 Conduct UQ/GSA simulations with improved uncertainty representations for LLNL gasoline surrogate model
5 Conduct UQ/GSA simulations for multi-component fuel blend FY16 – Q4
4
PROJECT APPROACH
Utilize ANL’s twin-piston RCM to acquire autoignition data
Rapid Compression Machine
Employ novel data analysis tools and advanced diagnostics– Physics-based, reduced-order system model;– Developing new diagnostics capabilities to better probe chemistry.
Synergistically improve kinetic models using novel analysis techniques(e.g., UQ/GSA) and detailed calculations/measurements of sensitiveprocesses (e.g., individual reaction rates)
6
TECHNICAL ACCOMPLISHMENTS / PROGRESS
Modifications and upgrades implemented since FY 2015 AMR to reduce measurement uncertainties and increase throughput– Redesigned heating system / control for mixture feedline from supply tank;– Continuous monitoring of reaction chamber temperature distribution;– High-resolution static pressure transducers (± 0.25 % of reading);– High-speed DAQ card (1 MHz, 16-bit) to improve heat release analysis
during high-temperature heat release;– Incorporated high-precision, automated feed valve for reaction chamber ;– Previously limited to ~16 shots/day; up to 40 shots/day now possible, with
excellent repeatability.
Twin-Piston Rapid Compression Machine
Operational challenges still exist– Synchronous twin-piston operation difficult to achieve; typically ~2-3 ms
delay. Efforts underway in FY 2016 to improve this, including integration of piston position sensor for high accuracy dynamic measurement;
– Piston seals often subjected to repeated thermal cycling, requiring replacement after 1-2 weeks of use. Different ring-pack configurations / seal designs and materials are being investigated.
TECHNICAL ACCOMPLISHMENTS / PROGRESS
Predictive modeling of LTC needed to guide design– Gasoline is complex, compositionally variant→ How do these features affect LTC behavior, especially
autoignition phenomena at low and intermediate temperatures (T = 600–1100 K)?
→ How can real fuels be represented by multiple-component (3-10) formulations?
→ Data are needed to compare autoignition behavior of real, full boiling range fuels with surrogates, including individual components, blends of these and mixtures with ethanol.
Investigating Gasoline and Surrogates http://www.sigmaaldrich.com/catalog/product/supelco/g006346?lang=en®ion=US
Commercial Gasoline Composition
9
TECHNICAL ACCOMPLISHMENTS / PROGRESS
5-member ring naphthenes– Acquiring combustion data for representative hydrocarbons at LTC
conditions. FY 2015/16 tests with CP, MCP and ECP;– Influence of ring substitution, temperature, pressure and dilution.
Task 1 – Surrogate Components
increasingdilution
increasingdilution
10
TECHNICAL ACCOMPLISHMENTS / PROGRESS
5-member ring naphthenes– Acquiring combustion data for representative hydrocarbons at LTC
conditions. FY 2015/16 tests with CP, MCP and ECP;– Influence of ring substitution, temperature, pressure and dilution.
Task 1 – Surrogate Components
11
TECHNICAL ACCOMPLISHMENTS / PROGRESS
5-member ring naphthenes– Acquiring combustion data for representative hydrocarbons at LTC
conditions. FY 2015/16 tests with CP, MCP and ECP;– Influence of ring substitution, temperature, pressure and dilution.
Task 1 – Surrogate Components
Pressure and dilution effects are much more substantial in NTC regime,while ring substitution significantly enhances low temperature chemistry.
12
TECHNICAL ACCOMPLISHMENTS / PROGRESS
FACE-F used as representative gasoline– Composition, properties well-characterized;– Investigating influence of ethanol blending levels (E0, E10, E20, E30, E100)
on autoignition chemistry; quantifying τ, LTHR/ITHR.
Task 2 – Gasoline / Ethanol Blends
13
TECHNICAL ACCOMPLISHMENTS / PROGRESS
FACE-F used as representative gasoline– Composition, properties well-characterized;– Investigating influence of ethanol blending levels (E0, E10, E20, E30, E100)
on autoignition chemistry; quantifying τ, LTHR/ITHR.
Task 2 – Gasoline / Ethanol Blends
Detailed kinetic model appears to give reasonable agreement.How can we infer performance under engine, e.g., HCCI, conditions?
14
TECHNICAL ACCOMPLISHMENTS / PROGRESS
Challenges exist comparing predictions of chemical kinetic models to LTC engine measurements– Many parameters in experiment (e.g., IVC) are estimated / uncertain, while
in-cylinder charge contains a distribution of temperature and composition;– Autoignition / heat release can proceed through various modes of
combustion, including sequential autoignition and deflagrative-propagation;– Uncertainties are inherent in chemical kinetic models due to uncertainties
in rates for elementary (and lumped) reaction steps• Higher model uncertainties at lower IVC temperatures and retarded combustion
phasing (as demonstrated in FY 2015 AMR presentation)
New approaches developed in FY 2016 to demonstrate correlation between RCM measurements and observed engine trends, but with much lower experimental uncertainties than engine experiments– Measurement campaign studied changes to required intake temperature in
order to maintain constant combustion phasing during HCCI combustion as intake charge is boosted, e.g., SAE 2013-01-2627
Task 2 – Gasoline / Ethanol Blends
15
TECHNICAL ACCOMPLISHMENTS / PROGRESS
HCCI data (UC Berkeley) covering Pin = 0.9 to 1.6 bar, Tin = 120 to 60 C RCM experiments conducted over well-resolved T, P conditions to match
those experienced by in-cylinder charge near end of piston compression– Pc = 15–80 bar, Tc = 800–1000 K at fuel loading of φ = 0.3, 21% O2
Task 2 – Gasoline / Ethanol Blends
16
TECHNICAL ACCOMPLISHMENTS / PROGRESS
HCCI data (UC Berkeley) covering Pin = 0.9 to 1.6 bar, Tin = 120 to 60 C RCM experiments conducted over well-resolved T, P conditions to match
those experienced by in-cylinder charge near end of piston compression– Pc = 15–80 bar, Tc = 800–1000 K at fuel loading of φ = 0.3, 21% O2
Task 2 – Gasoline / Ethanol Blends
T, P conditions during critical part of engine cycle captured by RCM measurements.Constant phasing trend (CA50 = 6° aTDC) can be observed on τ-Tc-Pc surface.
17
TECHNICAL ACCOMPLISHMENTS / PROGRESS
Constant combustion phasing in RCM indicated by iso-τ line. T, P at (τ = 4 ms) are close to those seen at TDC in UC Berkeley HCCI engine, trend with Tc–Pc very similar to that observed in Tin–Pin engine data.
Influence of NTC chemistry and appearance of LTHR clearly evident.
Task 2 – Gasoline / Ethanol Blends
τ = 4 ms
Appearanceof LTHR
Entering NTC
increasingpressure
τ = 4 ms
18
TECHNICAL ACCOMPLISHMENTS / PROGRESS
Ethanol-blended gasoline exhibits suppressed reactivity, most prominently at high pressure and low temperature, e.g., in boosted LTC range. This is consistent with observations of Pc = 20, 40 bar experiments.
Task 2 – Gasoline / Ethanol Blends
τ = 4 ms
increasingpressure
τ = 4 ms
19
TECHNICAL ACCOMPLISHMENTS / PROGRESS
Predicted ignition delay times for LLNL 5-component gasoline surrogate model are reasonably good, but evolution of low temperature (i.e., peroxide) chemistry, and associated LTHR at higher pressures not adequately captured.
Task 2 – Gasoline / Ethanol Blends
τ = 4 ms
Opportunity for improvement
increasingpressure
20
TECHNICAL ACCOMPLISHMENTS / PROGRESS
Predicted ignition delay times for LLNL 5-component gasoline surrogate model are reasonably good, but evolution of low temperature (i.e., peroxide) chemistry, and associated LTHR at higher pressures not adequately captured.
Task 2 – Gasoline / Ethanol Blends
τ = 4 ms
Opportunity for improvement
increasingpressure
Constant combustion phasing perspective highlights correlation between RCMdata, HCCI engine measurements; identifies opportunities for model improvement.
21
TECHNICAL ACCOMPLISHMENTS / PROGRESS
Methodology developed to implement UQ/GSA with LLNL model– ~104 correlated uncertainties. Apply UFs based on: (a) foundational (C0-C4)
chemistry (AramcoMech), or (b) reaction class (where ‘rate rules’ are used)– Detailed accounting of 7,366 reactions necessitates automated means to
identify, classify and appropriately perturb individual reactions
Task 4 – UQ/GSA for Gasoline Surrogate Model
LTC fueloxidation pathways
22
TECHNICAL ACCOMPLISHMENTS / PROGRESS
Efforts in FY 2016 focused on further development / refinement of approach and conditions considered
Conducted detailed review of uncertainties in foundational chemistry model (AramcoMech, 253 species, 1542 reactions)– Previously estimated with conservatively low uncertainty factors (UF = 2);– Realistic bounds are now considered for each reaction.
Refined techniques to account for ‘correlations’ within the model– Application to ‘rate rules’ for analogous reaction steps;– ‘Branching ratios’ that specify reaction rates with multiple product channels.
Task 4 – UQ/GSA for Gasoline Surrogate Model
iso-octane/air, 20 bar
RO2 = QOOH
iso-octane/air20 bar, 725 K
RH+OH = R+H2O combined
23
RESPONSE TO REVIEWER COMMENTS
The project has developed an excellent approach to overcome technical barriers of inadequate chemical kinetic modeling capability for LTC, and combustion data are acquired at pressures closer to actual in-cylinder conditions, which is important. The project is well designed in terms of systematic experiments and analysis for various fuels and surrogate blends and the approach seems feasible. The success of this endeavor will surely benefit the entire engine community. Furthermore, the RCM Workshop is an excellent way to engage labs, universities, energy companies and engine manufacturers. The project needs to continue to apply techniques and results to engine level attributes, covering more surrogate components and real fuels, while it would be appropriate to be more forward thinking and examine performance of other gasoline blends, such as gasoline with butanol.
– New experimental measurements are reported this FY, with efforts that explicitly illustrate correlations between constant volume conditions in the RCM, and variable volume conditions seen in engines, critically at engine-representative pressures. Plans for future fiscal years cover more gasoline types and surrogate formulations, while efforts in the Co-Optima program cover next-generation biofuels.
UQ/GSA tools aid chemistry development, while challenges and improvements in terms of accounting for correlated uncertainties are properly identified. The technical accomplishments are clearly aligned to the overall project objectives of improving chemical kinetic modeling capability, and the overall U.S. DOE goal of predictive engine simulations. It is unclear however, why ignition delay time is used exclusively as the metric for mechanism validation / improvement, when other measurements such as laminar flame speed, extinction strain rate, etc. should be considered. It would be beneficial to clearly identify how the kinetic model will be improved based on the measurements.
– Parameters derived from RCM measurements, such as ignition delay time, LTHR/ITHR and intermediate species concentrations are just a few that should be considered for kinetic model evaluation, and iterative improvement through sub-mechanism refinement. It should be noted that metrics such as flame speed are important for SI combustion, but are often dominated by small-molecule chemistry, and not as significantly influenced by the molecular structure / composition of the fuel.
COLLABORATIONS
DOE Working Groups on LTC engines: share data at meetings of AEC MOU, participate on ACEC Fuels Roadmap Subteam
CRC FACE Working Group: participate in meetings, test CRC fuels
ANL: gasoline LTC engine via FT045, refine UQ/GSA approaches and target reactions for mechanism improvement
LLNL: gasoline model development / validation, formulation of gasoline surrogates, ToolKit development / testing
KAUST / Chevron: provide FACE fuels, cyclopentane mechanism development
UC Berkeley, SNL: share HCCI engine data, tests with surrogate molecules
Northeastern U.: utilize NU’s novel mechanism diagnostics / analysis
Other organizations: NUI Galway (kinetic models); Vrije Universiteit Brussel (reduced-order physical models); U. Leeds (UQ/GSA approaches); U. Michigan / U. Connecticut / UL1ST / U. Cambridge (RCM review paper)
Ongoing Interactions (Inside / Outside VTO)
25
COLLABORATION
ANL-led, International RCM Workshop: patterned after ECN to better understand LTC phenomena using RCMs, especially autoignitionchemistry, turbulence-chemistry interactions, etc.– Participation includes experimentalists, modelers, theoreticians– Establishing consensus for ‘Best Practices’→ Approaches for reporting / analyzing / comparing data→ Approaches for simulating the experiments→ Uncertainty quantification for experiments and modeling
– 14 RCM laboratories contributed to first standardized tests using iso-octane; 3rd Workshop to be held 29 July 2016 at Yonsei University, Seoul, KOREA (in conjunction with International Combustion Symposium)
Community-wide Activities
26
REMAINING CHALLENGES / BARRIERS
Improvements to gasoline surrogate model requires deeper understanding of mechanism behavior, uncertainties associated with low temperature chemistry pathways of base model;
Understanding and representing the autoignition characteristics and sensitization of full boiling range gasoline via multiple-component (3-10) surrogate blends requires improved capabilities to formulate surrogates, wider palette of surrogate components;
Ignition delay time and preliminary heat release are integrated metrics for ignition chemistry, constraints exist with their utility; additional diagnostics, like direct measurement of important chemical intermediates, could improve development / validation efforts;
Mild ignition, which is initiated due to local perturbations in the gas temperature, can complicate autoignition measurements under some conditions and needs further understanding, as well as approaches to mitigate this.
27
PROPOSED FUTURE WORK
Coordinate surrogate blend formulation for FACE-F with LLNL and conduct tests with multiple-component surrogates– Utilize various techniques to select component molecules and blending
ratios, including blends with ethanol;– Conduct additional tests with surrogate components and full boiling-range
gasoline as needed.
Conduct RCM tests with RD-587 (E10 certification gasoline)– Coordinate with SNL to target LTGC engine conditions– Coordinate with LLNL to formulate and test surrogate blends for RD-587,
along with single-component fuels as necessary
Design, fabricate gas sampling / integrated analytical unit (GCxGC) for new single-piston RCM– Conduct detailed measurements with iso-octane and other surrogates at
LTC conditions to quantify intermediate species concentrations of important cyclic ethers and olefins formed / consumed during ignition delay period
– Collaborate with LLNL reduce uncertainty in rate controlling reactions and improve model predictions
FY 2016 and beyond
28
PROPOSED FUTURE WORK
Conduct tests with unique surrogate molecules, e.g., cycloheptadiene, in collaboration with Combustion Chemistry group at SNL CRF– Quantify influence of peroxide intermediates, e.g. QOOH, which have
been directly measured by CRF, on autoignition behavior at engine-relevant conditions (T, P, φ, %O2) via RCM measurements and kinetic modeling
FY 2016 and beyond
29
SUMMARY
Objective– Acquire data, validate / improve models for transportation-relevant fuels
at advanced engine conditions
Project Approach– Utilize ANL’s RCM and novel analysis tools, leverage expertise of BES-
funded researchers to synergistically improve predictive models
Technical Accomplishments / Progress– Acquired data to understand autoignition behavior of 5-member ring
naphthenes and FACE-F gasoline blended with ethanol– Improved / refined UQ/GSA using LLNL detailed gasoline surrogate model
Collaborations– National labs, universities and industry; International RCM Workshop
Future Work– Additional testing with gasoline surrogate components, blends and full
boiling range gasolines– Improvements to UQ/GSA, covering additional conditions / fuels